Professors awarded $1.1M by Keck Foundation to pursue brain research
Interdisciplinary endeavors have long been embedded within the Carnegie Mellon tradition of collaborative study. Cognitive neuroscience professor Marcel Just and computer science professor Tom M. Mitchell perpetuate this tradition with their groundbreaking efforts in brain research. The prestigious W.M. Keck Foundation awarded the duo a $1.1 million grant to carry out extensive research in the field of brain imaging.
The study, titled “Using fMRI Brain Activation to Identify Cognitive States Associated with Perception of Tools and Dwellings,” was a joint venture between the departments of neurology and computer science and was published on January 2 in PLoS one, an open-access, online publication by the Public Library of Science.
Just, who is also the co-director of the Brain Imaging Research Center at Carnegie Mellon, said that this is the first time that anybody has been able to track the specific object of a person’s thoughts.
“Previously, the best [that has been done] was to identify the category of objects, tools, or a type of dwelling that was being thought of,” he said.
Just, Mitchell, and their team employed a unique combination of brain scans and machine-learning algorithms to ascertain what a person is thinking of at a particular moment. Machine learning refers to a support system of artificial intelligence that uses algorithmic programs to help computers index information. This latest breakthrough can help form the relationship between a cognitive situation, such as the thought of an apple, and the activity behind this thought — the magnetic signals that are emitted when the thought of an apple occurs — in single and multiple brain regions.
The complexity of the human brain explicates itself through ongoing activity in that part of the central nervous system. This brain activity was measured by functional magnetic resonance imaging (fMRI) scanning.
While conventional magnetic resonance imaging (MRI) of the brain provides detailed imagery of the tissues, fMRI shows the magnetic energy emanating from the brain.
The researchers used the Siemens Allegra 3.0T scanner, an innovative machine at the collaborative Brain Imaging Research Center shared by Carnegie Mellon University and the University of Pittsburgh.
“We’ve learned how the brain represents concepts that we describe by concrete words,” Mitchell stated in a Carnegie Mellon press release.
The brain activity levels, in terms of magnetic signals, can be assessed using oxygen levels in different sections of the brain. When there is a greater amount of oxygen in the active region of the brain, a stronger magnetic signal is generated. A stronger magnetic signal suggests higher neural activity.
After obtaining a trend of magnetic signals, the team applied computer algorithms to these scanned images in order to decipher the signals transmitted during brain activity.
The first phase of applying algorithms involves finding voxels that express a similar pattern over a number of trials. Voxels are volume elements normally used in the examination of medical data. While pixels characterize 2D image data, voxels represent 3D image data. The second phase focuses on subsets of voxels. Each voxel can be virtually anywhere in the many possible levels of activity in the brain.
These algorithms help scientists acquire the mapping between individual items and brain activity levels.
In the context of the published study, the team tested algorithms on a group of subjects and discovered that the codes emitted in most human brains are similar.
As of yet, this technology only applies to concrete objects, like apples. However, according to Just, in the future, more abstract notions, such as people or ideas, will also be explored through this examination of the human brain.
Just and Mitchell intend to use a variety of methods that will help interpret brain activity in terms of language, according to the press release. Computational studies of English text, along with fMRI scans of the brain interpreting words, activating signals, and translating those signals into concrete ideas largely shape the team’s future plans.
Despite the advanced nature of brain imaging, students play an important role in this research, Just said. From student participants to those who have completed their honors thesis in areas of neuroscience, there is significant student involvement in the field.
Down the line, the professors’ insight into the brain activity may help gain a better understanding of neurological disorders such as schizophrenia and dyslexia.